Predicting social adjustment in middle childhood: the role...

Articles

Predicting social adjustment in middlechildhood: the role of preschool attachmentsecurity and maternal styleCathryn L. Booth, University of Washington, Linda Rose-Krasnor,Brock University, Jo-Anne McKinnon and Kenneth H. Rubin,University of Waterloo

Abstract

Children's social and emotional adjustment at age 8 were examined in relation toattachment security, parenting style, setting conditions, and social and emotionaladjustment at age 4. Seventy-nine children participated in videotaped interaction ses-sions with their mothers and with unfamiliar peers at the two ages. Data werederived from videotape coding, mother questionnaires, and child sociometric ratings.Results indicted that internalizing problems, externalizing problems, and socialengagement were related at the two ages. After removing the variance due to therelationship between child behaviors at the two ages, a comparison of mother-childrelationship predictors indicated that attachment security at age 4 was the strongestpredictor of internalizing problems and social engagement/acceptance at age 8, whilematernal style was the strongest predictor of externalizing difficulties. Results pointto the importance of both aspects of the mother-child relationship, and indicate thatthe nature of family and peer links may vary depending upon the specific socialdomain assessed.

Keywords: Attachment; peers; parenting.

How children Ieam to initiate and maintain positive relationships with others arecritical questions in the study of child development. The significance of the chil-dren's relationships in the peer group, for example, is highlighted by researchindicating that peer rejection and social isolation are associated contemporane-ously and predictively with indices of psychological maladjustment (see Parker &Asher, 1987).

Although the proximal causes of poor peer relationships have received a gooddeal of attention there are fewer data that address the origins and maintenance ofThe research reported herein was supported by grant #NR01635 from the National Center forNursing Research, and grant #HD27806 from the National Institute of Child Health and HumanDevelopment. We gratefully acknowledge the contributions of Patricia Taylor, Donna Dwyer-O'Connor, Marilyn Dryden, Sheila Meagher, Shannon Gird, Susan Coates, and Sybil Evans(University of Washington) and Cherami Wichmann, Loretta Lapa, Kerri Hogg, and Robert Coplan(University of Waterloo) for their help in data collection, coding, and analysis. We thank the mothersand children who made this project possible. Requests for reprints should be sent to Cathryn L.Booth, Child Development and Mental Retardation Center, Mail Stop WJ-10, University ofWashington, Seattle, WA, 98195.

© Basil Blackwell Ltd. 1994. Published by Blackwell Publishers, 108 Cowley Road, Oxford OX4 lJF, UK and 238 Main Street,Cambridge, MA 02142, USA.

Cathryn Booth, Linda Rose-Krasnor, Jo-Anne McKinnon and Kenneth Rubin

positive and negative relationships in the peer group. For example, children whoare aggressive or anxious/withdrawn appear to be rejected by their peers, while chil-dren who display competent social behaviors are generally accepted by their peergroup. But the origins of these behavior profiles are relatively unknown. One signif-icant factor is undoubtedly the nature of the parent-child relationship (e.g. Sroufe& Waters, 1977; Putallaz, 1987; Parke, MacDonald, Beitel & Bhavnagri, 1988;Dodge, Bates & Pettit, 1990; Ladd, 1991; Booth, Rose-Krasnor & Rubin, 1991).

The mother-child relationship and the mother's parenting behaviors are com-plex phenomena that can be studied from a variety of perspectives, only some ofwhich may be relevant to the child's developing and maintaining positive relation-ships outside the family unit. One such perspective stresses the importance of themother-child attachment relationship. We have shown (Booth et al, 1991), ashave others (e.g., Lieberman, 1977; Pastor, 1981; Sroufe, 1983; Youngblade &Belsky, 1992), that a history of insecure attachment to the mother in infancy ispredictive of social maladaptation with peers in the preschool years. It has alsobeen demonstrated that insecurity of attachment during the preschool period isconcurrently and predictively associated with lack of social competence and mal-adaptive peer relationships (e.g., Lewis, Feiring, McGuffog & Jaskir, 1984; Park& Waters, 1989; Turner, 1991).

The connection between the quality of the mother-child attachment relationshipand the child's social competence has been conceptualized in terms of the childdeveloping, through the primary attachment relationship, an internal workingmodel of the self in relation to others (Bowlby, 1973; Bretherton, 1985; Main,Kaplan & Cassidy, 1985). That is, through continuing transactions with theattachment figure, the child develops complex mental representations, or internalworking models, of the attachment figure and of the relationship between self andother. These models then help guide the child's behavior with the attachment fig-ure and with others.

Viewing internal working models from an emotional, rather than a cognitiveperspective, the child who believes that his/her parent is available and responsiveto his/her needs will feel secure, confident, worthy of love, and self-assured whenintroduced to novel settings. Felt security is thought to result in active explorationof the social environment. In turn, exploration of the social environment leads topeer play; these play experiences, noted above, are likely to result in the develop-ment of social competencies. On the other hand, the child who believes that his orher parent is unavailable and unresponsive will feel insecure, unworthy of love,and wary in novel situations. These feelings of insecurity would be expected tohinder exploration of the social environment, diminishing the opportunities forpeer play and perhaps hindering the consequent development of social competen-cies with peers.

Ultimately, the chain of events described above leads to the child interacting inadaptive or maladaptive ways with others. Consistently maladaptive interactionswith peers are likely to have the following consequences: (a) the child will recog-nize his/her social failure in the peer group and thus develop negative internalworking models of the self in relation to peers; (b) the child will have difficultydeveloping intimate relationships with age-mates (i.e., friendships) because of feel-ings of insecurity, distrust, and self-perceptions of social incompetence; (c) thechild will solidify her/his negative peer reputation by isolating her/himself fromthe peer group (social withdrawal) or by reacting in anger to the peer group

© Basil BiackweU Ltd, 1994 Social Development, 3, 3, 1994

Predicting Social Adjustment 191

(aggression). These consequences do not auger well for a prognosis of well-being.In addition to the influence of the mother-child attachment relationship on chil-

dren's well-being, parenting behaviors or styles appear to wield significant influ-ence. The means by which parents manage and discipline their children, and theaffective context within which parenting behaviors are emitted, have long beenthought to play a crucial role in the development of psychological well-being andmaladjustment.

For example, parents who are relatively cold and rejecting, physically punitive,and who discipline their children in an inconsistent manner are more likely tohave aggressive, psychologically undercontrolled children than are those parentswho are relatively warm, supportive, and authoritative (see Pepler & Rubin,1991). Parents who are overcontroUing and restrictive, and who actively encour-age child dependency, are thought to have children whose emotions and behaviorsare psychologically overcontroUed (see Mills & Rubin, 1993). Finally, parentswho have reasonable expectations for mature behavior, who use control anddemands appropriately when their children are disobedient or display deviantbehavior, and whose parenting behaviors are emitted in an affectively warm andnurturant environment are thought to have psychologically well-adjusted children(Maccoby & Martin, 1983).

Much of the research on parenting styles has not focused specifically on theeffects of different styles on children's peer relationships. However, it is clear thatthe adaptive or maladaptive child behaviors that are related to differences in par-enting would be expected to affect the child's relationships both within and out-side the family context.

To summarize, from the perspective of attachment theory, the child is viewedas developing a cognitive representation of relationships that guides him/her insubsequent interactions with peers. These representations develop over time, onthe basis of the quality of early maternal care. From the perspective of parentingstyle, the child is viewed as developing patterns of behavior in response toparental behavior, and these adaptive or maladaptive behavior patterns affect thedevelopment and quality of peer relationships.

On the surface, the separation of attachment and parenting influences mightappear to be an artificial one: We would expect children's attachment security tobe related to parenting styles, in the sense that most mothers who provide a sensi-tive responsive early environment that is conducive to the development of thechild's attachment security, would be expected to provide a warm nurturing envi-ronment and use appropriate control techniques when the child is older. However,some mothers may not be as consistent over time in the care that they provide,due to changes in their own lives, or their lack of skills or motivation to changetheir parenting behaviors in response to changes in the child. For example, somemothers may be adept at nurturing a young infant, but may react inappropriatelyto the child's later need for autonomy.

Attachment theory, strictly interpreted, would emphasize the child's early moth-ering as the source of internal working models of relationships, which are viewedas self-reinforcing and decreasingly malleable with development. The parentingstyles approach would emphasize the continuing influence of the parent's behavioron the child. Thus, we might expect a statistical relationship between attachmentsecurity and parenting styles, but attachment theory would predict that the child'ssecurity would be more closely related to social and emotional outcomes.

© Basil Blaekwell Ltd. 1994 Social Developtnent, 3, 3, 1994

192 Cathryn Booth, Linda Rose-Krasnor, Jo-Anne McKinnon and Kenneth Rubin

More distal aspects of the family environment can also afTect the child. Suchsetting conditions may be indexed by demographic marker variables, or prefer-ably, factors linked more directly (albeit still distally) to the quality of themother-child relationship, namely, variables such as maternal stress, depression,and the availability of social support (see Rubin & LoUis, 1988). For example, anumber of studies have found that maternal stress is negatively related to thequality of mother-child interaction and attachment security (e.g., Cmic &Greenberg, 1990; McKinnon, Rubin, Booth & Rose-Krasnor, 1993; Teti,Nakagawa, Das &Wirth, 1991). Maternal mental health has been linked to thequality of the mother-child attachment relationship (e.g., Spieker & Booth, 1988;Radke-Yarrow, Cummings, Kuczynski & Chapman, 1985). Finally, the availabil-ity of social support is negatively related to maternal restrictiveness and punitive-ness and positively related to maternal responsivity and sensitivity and the qualityof the mother-child relationship (e.g., Rose-Krasnor, Rubin & Booth, 1993;Weinraub &Wolf, 1987; Crockenberg, 1981). Taken together, we would expectthat setting conditions would most likely afTect the mother's ability to interactwith her child in positive ways, thus having a direct effect on parenting style, butan indirect effect on the child's social and emotional adjustment.

In the ongoing longitudinal study that is the subject of this report, we foundthat children's social competence with peers at age 4 was related to concurrentmeasures of parenting style and to the children's attachment security to mother(Rose-Krasnor, Rubin, Booth & Coplan, 1994). The main purpose of the currentreport was to (a) evaluate the continuities in social and emotional adjustmentbetween ages 4 and 8, and (b) to compare setting conditions, attachment securityand maternal style at age 4 in terms of their prediction of social and emotionaladjustment at age 8. Although other studies of family and peer links have focusedon attachment quality, or maternal style, or other aspects of the mother-childrelationship, the present study is relatively unique in its emphasis on a comparisonof mother-child relationship influences on child social and emotional adjustment.

We hypothesized that there would be continuity in social-emotional adjustmentfrom age 4 to 8. Additionally, we hypothesized that age-4 attachment security (asa measure of the child's internal working model of the self in relation to mother)would be a better predictor than setting conditions or maternal style of age-8social and emotional outcomes. This hypothesis is derived directly from attach-ment theory, in the sense that family and peer relationships are viewed as beinglinked via the process of the child developing an internal working model of rela-tionships, rather than the child exhibiting specific social behaviors based onmaternal modeling, or based on the general climate of stress and support in thehome environment.

Methods

Overview

During the first phase of this longitudinal research project, mothers and theirfour-year-old children participated in a mother-child-peer interaction session.Unfamiliar, same-sex, same-age (± 3 mos) children were assigned to dyads on thebasis of attachment classiflcations at age 20 months or at age 4 years (assessedprior to the mother-child-peer session). The dyads consisted of a "control' child

© BasU BiadcweU Ltd. 1994 Social Development, 3, 3, 1994


who was securely attached, and a "focal" child who was either securely- or inse-curely-attached to his/her mother. The focal child's mother also participated.Measures were obtained of setting conditions, maternal interactive style, childattachment security, and child social and emotional adjustment in the areas ofsocial competence with peers and behavior problems.

At age 8, each focal child returned to the laboratory for a peer-quartet session.Children were placed in quartets with same age (± 6 mos), same-sex playmateswho were recruited via posters placed in schools, community centers, etc. Allplaymate children met the criterion of behavior problem scores below the clinicalcut-off on the Child Behavior Checklist (Achenbach, 1991). As before, measureswere obtained of child social and emotional adjustment in the areas of socialcompetence and behavior problems.

Subjects

The subjects were 79 children and their mothers. As of the 8-year assessment, themothers' mean age was 35.67 (± 5.0) years; 13% had competed high school, 38%had completed some college, 30% were college graduates, and 19% had anadvanced degree. Social status, as measured by the HoUingshead (1975) Four-Factor Index, averaged 44.2 (± 12.0). Most of the mothers (81%) were married.

The children were Caucasian primarily (87%); 54% were male and 61% werefirst-bom. The children were assessed at 4.3 (± .1) years and again at 8.0 (± .15)years.

Procedures

Age 4. The focal and control children and their mothers were videotaped interact-ing in a laboratory playroom, in a wide variety of structured and unstructured sit-uations, as follows (a) Mother-Child Free Play. Focal child and mother interactedin the playroom with toys (10 mins); (b) Free-Play Quartet. Focal and control chil-dren and mothers interacted (10 mins); (c) Block-Building: Control child, and focalchild and mother built a house out of Duplo blocks (10 mins); (d) Free-Play Triad:Control child, and focal child and mother interacted in the playroom with toys (10mins); (e) Separation/Novel Toy: Focal mother left the room; one novel action fig-ure was introduced (10 minutes); (f) Cookie Sharing: The children were given onelarge cookie to share (5 mins); (g) Reunion: Focal mother returned (5 mins).

Mothers also completed questionnaires about demographics, social support,depression, stress, and child behavior problems prior to the lab visit.

Age 8. At age 8, the focal children returned to a larger laboratory playroom forquartet interaction with a new set of unfamiliar peers. The observational para-digm comprised several segments including a free-play session (20 mins); a sessionduring which each child was required to sort tickets cooperatively (10 mins); anda session during which a single remote controlled car was introduced to the play-room (20 mins). The playroom contained a variety of age-appropriate toys andgames: Barbie dolls and Barbie "equipment"; Teenage Mutant Ninja Turtles and"equipment," board games ("Sorry" and "Trouble"), large cardboard buildingblocks, drawing paper, markers, and books.

At the end of the session, each child was interviewed separately, and confiden-

© Basa BiackweU Ltd. 1994 Social Development, 3, 3, 1994

194 Cathryn Booth, Linda Rose-Kramor. Jo-Anne McKirmon and Kenneth Rubin

tially, to obtain sociometric information about the other members of the quartet.Each child was asked the following set of five questions about the other quartetmembers individually, using photographs taken at the beginning of the laboratorysession: (a) "How much did you like to play with (child's name)?", (b) "Howmuch did you like to work with (child's name)?", (c) "How much did (child'sname) cooperate or help?", (d) "How much did (child's name) argue or fight?",and (e) "How shy was (child's name)?". A five-point rating scale (from "not atall" to "a lot") was used for each question. The scores for a, b, & c were highlyintercorrelated and therefore were averaged to produce a "likeability" index.

Observational Coding

Four observational coding schemes were used to code the videotaped data:Maternal Warmth and Control Rating Scales (Rubin & McKinnon, 1993) to codecontrol and warmth during block-building, triad free play and mother-child freeplay at age 4; the Reunion Rating Scale (Booth &Perman, 1989) to evaluate thechild's security of attachment to mother at age 4; the Play Observation Scale(Rubin, 1989), to code the child's behavior during the novel toy segment at age 4,and during free play at age 8; and the Ticket-Sorting Coding Scheme (Coplan,Rubin, Fox, Calkins & Stewart, 1994), to assess the child's disruptive behaviorduring this task at age 8. Data were coded independently for each coding scheme,by separate sets of blind observers.

Maternal Warmth and Control Rating Scales. These scales (Rubin &McKinnon, 1993) consisted of a series of 3-point ratings in the areas of proxim-ity, positive affect, responsivity, positive control/guidance, hostile affect, negativeaffect, and negative control. Ratings of the videotaped interactions were com-pleted after every minute of interaction during the Mother-Child Free Play,Block-Building, and Free-Play Triad segments, and averaged across minutes. Forthe observation segments that included both children, only interactions betweenthe mother and her own child were rated. Interobserver agreement on the scalesranged from .86 to .97, calculated on 10% of the cases.

Reunion Rating Scale. The Reunion Rating Scale (RRS; Booth & Perman,1989) was used to evaluate each focal child's security of attachment to mother atage 4. This 5-point scale rated the child's security as demonstrated upon reunionwith the mother (following the cookie-sharing segment). The scale was developedto provide a quantitative measure of degree of security, which was of greaterinterest in this context than were the typical A-B-C-D classifications.Additionally, the scale provided contemporaneous assessment of attachment secu-rity for all subjects (Recall that half of the subjects had 20-month classificationsfrom the Strange Situation and half had 4-year classifications from a mother-childlaboratory visit prior to the peer session).

The RRS was based on the preschool attachment classification and rating sys-tem developed by Cassidy and Marvin with the MacArthur Working Group(1989). Coders who rated security in the present study also participated in theMacArthur Working Group and were trained and certified on the Cassidy andMarvin classification and rating system. The RRS is similar to the Cassidy andMarvin security rating scale that is used to assist the observer in making a classifi-cation. Interrater reliability on the RRS, computed for 12% of the subjects,yielded an intradass correlation coefficient of .74.

© Basil Blaekwell Ltd. 1994 Social Development, 3, 3, 1994


To evaluate the validity of the RRS, scores were compared for children classi-fied as secure or insecure by other means. Specifically, 4-year attachment classifi-cations (using the Cassidy, Marvin, et al, 1989 system) had been obtainedindependently on the nonlongitudinal subjects from a separation-reunion proce-dure approximately 1 month prior to the peer session; and 20-month attachmentclassifications (using the Ainsworth et al, 1978 system) had been obtained inde-pendently on the longitudinal subjects in the Strange Situation. The average RRSscore of children previously classified as secure at 4 years was 3.58, comparedwith 2.10 for the insecures, F (1,53) = 8.62, p < M. Similarly, the average RRSscore of 20-month secures was 3.85, compared with 2.81 for the insecures, F(1,51) = 7.88, p < .01.

Play Observation Scale. The Play Observation Scale coding taxonomy (POS;Rubin, 1989) was used to assess the child's social and nonsocial behavior duringvideotaped interaction with peers at ages 4 and 8. Data were coded every 10 sec-onds into one of the following social participation categories: onlooker behavior;unoccupied activity; and solitary, parallel, or group play. Nested within the latterthree categories were the cognitive play categories of functional-sensorimotor,constructive or dramatic play, and games-with-rules. Also, exploratory activity,and conversations in the absence of play were coded. Data were converted to pro-portions of total time intervals, and subjected to arcsine transformation. Cohen'skappa, on 20% of the videotapes, ranged from .70-.83 for social and cognitiveplay categories.

Ticket-Sorting Coding Scheme. This coding scheme (Coplan et al, 1994) wasused to assess the proportion of time spent in on- and off-task behavior duringthe 10-minute segment in which the children were asked to sort tickets coopera-tively. Behaviors were considered "off-task" if they did not include such actionsas picking up and placing the tickets in various piles or talking about the task athand. Off-task behavior was divided further into Unoccupied and Disruptive cate-gories (i.e., goofing off or disrupting others from the task at hand).

Measures

Setting conditions. Setting conditions were assessed via measures of maternalstress, social support, and depression. One stress measure was the LifeExperiences Survey (LES; Sarason, Johnson & Siegel, 1978), a 61-item question-naire that yields a weighted score for maternal perceptions of the impact of nega-tive life changes. The second stress measure was the Parenting Daily Hassles scale(PDH: Cmic & Greenberg, 1990), which is a 20-item questionnaire that assessesthe frequency and intensity of daily hassles due to typical everyday parenting situ-ations.

Social support was assessed with the Personal Resources Questionnaire (PRQ;Brandt & Weinert, 1981), a 25-item Likert-type scale that is based on Weiss's(1974) conceptualization of the functions of social relationships. The BeckDepression Inventory (BDI; Beck, 1970), a 21-item scale, was used to measure theseverity of depression. The total score refiects both the number and severity ofsymptoms.

A principal components analysis with varimax rotation on the summary scoresfrom the LES, PDH, PRQ, and BDI revealed that the LES, PDH, and BDI hadhigh positive loadings, and the PRQ had a high negative loading on one factor.

© Basil Blackwell Ltd. 1994 Social Development, 3, 3, 1994


Consequently, the factor score was used in all subsequent analyses to representnegative setting conditions.

Maternal style. Two aggregated scores were derived from the observationaldata - maternal warmth and maternal negativity. For maternal warmth, the sum-miary score consisted of summed z-scores for ratings of proximity, positive affect,responsiveness, and positive control. Internal consistency of the composite was.71. The maternal negativity score (alpha=.73) consisted of summed z-scores forratings of negative control, hostile affect, and negative affect.

Attachment security. The Reunion Rating Scale score was used as the measureof security of attachment to mother at age 4.

Child social and emotional adjustment - age 4. Measures of child social andemotional adjustment at age 4 were in the areas of (a) internalizing problems, (b)externalizing problems, (c) social engagement, and (d) reticence.

Internalizing and externalizing behavior problems were assessed by means ofthe Child Behavior Checklist (Achenbach, 1991), a 118-item parent-completedch«;klist designed to assess behavior- and social-competence problems in children.The internalizing and externalizing summary T-scores were used in the analyses.

The aggregated social engagement measure from the POS consisted of summed,arcsine-transformed proportions from the POS of positive or neutral conversationwith the peer, and positive or neutral play with the peer. This social engagementmeasure is construed as an important component of social competence. The reti-cence measure was the sum of arcsine-transformed proportions from the POS foronlooker- and unoccupied behavior.

Child social and emotional adjustment - age 8. As at age 4, the assessments atage 8 were designed to evaluate internalizing problems, externalizing problems,social engagement, and reticence. However, unlike the age-4 data, multiple assess-ments of each construct were completed at age 8. Consequently, factor scoreswere used for these variables at age 8. The advantage of using factor scoresinstead of single indicators is that the factors comprised measures obtained fromseveral informants, using a variety of methods.

Four factor scores were used as dependent variables. These scores were identi-fied on the basis of a series of principal-components analyses with varimax rota-tion. In all cases, factor loadings exceeded a minimum inclusion criterion of .40,and the factors were readily interpretable.

The internalizing- and reticence factor scores were orthogonal factors in oneanalysis. For the internalizing factor score, the variables with high loadings werethe POS proportions for solitary passive play (solitary-constructive + solitary-exploratory), and solitary active play (solitary-functional + solitary-dramatic);*and the internalizing T-score from the CBCL. For the reticence factor score, thevariables with high loadings were the proportion for observed reticence from thePOS (onlooker- and unoccupied behavior), and the shyness score from the socio-metric ratings.

The externalizing factor consisted of the proportion of disruptive behaviorobserved during the ticket-sorting task, the sociometric rating for aggression, andthe externalizing T-score from the CBCL. Finally, the social engagement andacceptance factor consisted of the relevant proportions from the POS coding (pos-itive or neutral conversation with the peer, and positive or neutral play with thepeer), and the likeability score from the sociometric ratings by the quartet-peers.

o BasU BiackweU Ltd. 1994 Social Development, 3, 3, 1994


Results

Descriptive statistics for all variables are shown in Table 1 and intercorrelationsamong them, in Table 2. The bivariate correlations indicated that age-4 measuresof child social and emotional adjustment were significantly related to the compa-rable age-8 measures, with the exception of reticence. The correlations alsorevealed that setting conditions, maternal warmth and negativity, and attachmentsecurity at age 4 were related both concurrently and predictively to measures ofsocial and emotional adjustment.

A series of multiple regression analyses was conducted to evaluate the relation-ship between age-4 predictors and age-8 social and emotional outcomes (i.e.,internalizing problems, externalizing problems, social engagement, reticence). Onthe first step of each analysis, the age-4 "matching" child variable was entered.For example, the CBCL internalizing score at age 4 was entered for the predictionof the internalizing factor score at age 8. On the second step, age-4 setting condi-tions, maternal warmth, maternal negativity, and attachment security wereentered as a block.

The results of the regression analyses are shown in Table 3. In the first analysis,internalizing problems at age 4 predicted internalizing problems at age 8. In thesecond block of variables, attachment security was the only one that added signif-icant variance to the prediction of internalizing problems (in a negative direction).

The pattern was different for externalizing problems. The significant predictorsin the externalizing analysis were externalizing problems and maternal warmth (ina negative direction) at age 4. None of the other predictors was significant.

For social engagement/acceptance at age 8, attachment security was the onlysignificant predictor following the entry of social engagement at age 4. Finally,none of the variables predicted reticence.

Discussion

The results supported the hypothesized continuities in the children's social andemotional adjustment from age 4 to age 8. The cross-age correlations were signifi-cant for internalizing problems, externalizing problems, and social engagement.Thus, despite differences in types of assessments, observation contexts, and playpartners at the two ages, indices of social and emotional adjustment were rela-tively stable.

Unlike the other variables, reticence was not related at the two ages. It isunclear why this would be the case, although the differences in observation con-texts at age 4 and age 8 might account for the lack of stability. That is, onlookerand unoccupied behavior in a dyad (at age 4) might be qualitatively differentfrom these same behaviors in a quartet (at age 8). Additionally, recent evidencesuggests that the various forms of nonsocial activity in the presence of peers maychange in their meaning and consequences with development (see Rubin &Asendorpf, 1993; Coplan et al, 1994).

We had hypothesized that attachment security at age 4 would be a better pre-dictor than setting conditions or maternal style, of social and emotional adjust-ment at age 8. This hypothesis was supported in part. The multiple-regressionanalyses indicated that attachment security (but neither setting conditions, normaternal warmth or negativity) contributed significantly to the prediction of



TaUe 1. Descriptiye Statistics for Variables Prior to Aggregation

Setting ConditionsLES - NegativeParenting Daily HasslesBeck Depression InventoryPersonal Resources Quest.

Matemal WarmthProximityPositive AffectResponsivenessPositive Control

Matemal NegativityNegative ControlHostile AffectNegative Affect

SecurityCBCL - InternalizingCBCL - ExternalizingReticenceSocial Engagement

Internalizing FactorCBCL - InternalizingSolitary ActiveSolitary Passive

Externalizing FactorCBCL - ExternalizingDisruptiveAggression Rating

Reticence FactorReticenceShyness Rating

Social Engagement/Acceptance FactorSocial EngagementHelpfulness RatingWork RatingPlay Rating

N

Age 4

79797979

73737373

7373737479797272

Age 8

797979

797979

7979

79797979

Mean

5.6640.56

5.58147.58

2.471.602.901.44

1.181.061.043.54

50.6051.66

.17

.56

50.54.06.01

49.38.01

1.14

.121.50

.343.983.853.93

SD

6.147.125.69

17.23

.26

.24

.22

.16

.19

.09

.051.449.529.41

.15

.20

9.26.13.03

9.95.04.29

.11

.54

.25

.74

.64

.63

internalizing problems and social engagement/acceptance at age 8, beyond whatwas predicted by, respectively, internalizing problems and social engagement atage 4.

These data correspond well with a growing body of research linking the parent-child attachment relationship to the development of social competence in children

Basil Blackwell Ltd. 1994 Social Development, 3, 3, 1994


9t

O

o

§1"3

H

e

M̂

I I

oo -"̂ '— —< t ^(N O ^ O —'

'-^ ON r-O O Nr

*OO

I I

* * •* • ** * * *O O fN '-^

^ c

C r?1 So

U U -B 6

<N

ro fN O

*

-.3

1

•**OS

1 vq

ro 00

o o

1

.08

.13

32**

**

-.50^

-.25

*-.

1632

**

.16

.37*

**33

**.1

2.4

6***

.25*

*

o

-.09

-.08

o

r

S

-.10

-.23

*-.

12.3

7***

00 O

*OO VO - - I

^ <N q

r r0 0

CO(N

<N(N

c .t 00

4J

bo

.3I o

V

*

qVI

TOS ^

^ V

© Basil Blackwell Ltd. 1994Social Development, 3, 3, 1994


Table 3. Regression Analyses Predicting Child Social and Emotional Outcomes atAge 8 from Age 4 Variable

Step

12

12

12

Variable

CBCL-Int.Security

CBCL-Ext.Warmth

Social Engage.Security

Beta

Internalizing.39

-.30

Externalizing.47

-.28

R

Problems.41.51

Problems.46.54

Social Engagement/Acceptance.20 .29.31 .42

R̂

.17

.26

.21

.30

.08

.17

F-to-Enter

14.35***8.32**

21.82***7.95**

3.21 +7.43**

Note. None of the variables predicted reticence.+ /7< .10; ** p< .01,*** p< .001.

(e.g., Cohn, 1990; Park «& Waters, 1989; Turner, 1991). Our findings are consis-tent with the notion that a child who is secure in his or her relationship with theprimary attachment figure has the confidence to explore unfamiliar social environ-ments and has positive expectations for interactions with others. The secure child,therefore, should show high levels of positive social play; the insecure child shouldlack social confidence, hold negative expectations for social outcomes, and thus,be less likely to engage in interactions with unfamiliar peers. Our data supportthese theoretically-based associations.

The prediction of externalizing problems at age 8 yielded different results:Maternal warmth, but not attachment security, was a significant (negative) predic-tor of externalizing problems at age 8. Although maternal warmth was thestrongest predictor, the bivariate correlations between maternal negativity andexternalizing problems, and between setting conditions and externalizing problemsalso were significant, in a positive direction. Attachment security was not relatedto externalizing problems.

These results suggest that externalizing behavior may be infiuenced moredirectly by the mother's interactive style, than by the child's internal workingmodels of relationships. That is, when interacting with peers, the child with exter-nalizing difficulties may be exhibiting a general reaction to the mother's relativelynegative behavior toward him/her, or the child may be imitating more specificsocial behavior modeled by the mother. This interpretation assumes a chain ofcausation from mother to child. Alternatively, it is possible that the mother'sbehavior may be a response to the child's externalizing tendencies, rather than acause.

The setting conditions variable at age 4 was not a significant predictor in anyof the multiple regression analyses. However, the bivariate correlations indicatedthat this variable was positively related to internalizing and externalizing prob-lems and negatively related to social engagement/acceptance at age 8. Theseresults are consistent with previous research indicating the negative effects on the

© Basil BiackweU Ltd. 1994 Social Development, 3, 3, 1994


child of maternal stress, depression, and poor social support. Yet, it is also thecase that setting conditions are relatively distal factors that probably affect thechild indirectly by means of their more direct effects on the mother's ability tointeract with her child in positive ways, and to develop a relationship that is opti-mal for the child. In the present study, the relative strength of the more proximalpredictors (i.e., attachment security and maternal style), in comparison with set-ting conditions, lends support to this interpretation.

It is noteworthy that security of attachment was positively correlated withmaternal warmth, and negatively correlated with maternal negativity at age 4.These results suggest relative continuity in maternal behavior, if we assume thatthe development of a secure child-mother attachment relationship is aconsequence of maternal sensitivity and responsivity in infancy. Yet, attachmentsecurity was a better predictor of internalizing problems and social engagement/acceptance than was maternal style, and maternal style was a better predictor ofexternalizing problems. These results indicate that in investigating the origins ofchild social behavior, it is important to examine the separate contributions of var-ious aspects of the parent-child relationship, despite the statistical connectednessof these variables. Recent evidence from other studies (Greenberg, Speltz &DyKlyen, 1993; Waters, Posada, Crowell & Lay, 1993) lends support to this view.

Others have argued that different types of attachment insecurity (avoidance,resistance, disorganization) should relate to varying forms of maladaptive behav-ior with peers (Sroufe, 1983; Erickson, Sroufe & Egelund, 1985; Cohn, 1990). Inthe present study, we did not have sufficient representation among the varioustypes of insecurity to perform such analyses. Further, we chose to measure secu-rity on an ordinal scale rather than to use the A-B-C-D classification system,because the amount of security or insecurity, rather than the type was moreimportant in the present context. Nonetheless, it is possible that analysis by typewould yield different results.

Also, it should be noted that these results were obtained by observing children'sinitial (i.e., one-occasion) interactions with unacquainted peers. If we hadobserved children with their best friends, or with an established peer group, theresults might differ. However, we would expect somewhat comparable results inthese other contexts, given that the behavior we observed in the novel-peer situa-tion was consistent with parental reports of behavior problems, and these reportspresumably reflect the children's behavior in a variety of relationships and con-texts. The advantage of the unacquainted-peer method is that it allowed us toevaluate each child's social competence and adaptive or maladaptive behavioroutside the context of established behavior patterns among acquainted peers. It isthese social competencies, or lack thereof, that play an important role in thedevelopment of positive or negative relationships with others, and, if negative,place the child at risk for subsequent psychological maladjustment.

It has been suggested that security of attachment to mother should be moststrongly linked to the quality of children's close relationships with friends (Park &Waters, 1989; Youngblade & Belsky, 1992). However, we have found that study-ing children's behavior with unacquainted peers allows us to tap a more primaryelement in the process of developing good-quality friendships - i.e., the socialcompetencies that form the basis for the development of intimate relationships.

The results of the present study add to, and support the growing literature onthe links between mother-child and peer relationships. Yet, this study is relatively



unique in its emphasis on comparison of mother-child relationship influences onmultiple aspects of child social and emotional adjustment. Additionally, this studyis one of the few longitudinal investigations in the area to employ comparablemeasures of social and emotional adjustment during preschool and middle child-hood; as such, the design of the study allowed us to evaluate the stability of chil-dren's social competencies and adjustment difficulties across these ages, and toseparate out the variance due to stability in child behavior from the variance dueto earlier mother-child relationship influences.

The purpose of this report was to compare the relative value of age-4 variablesin predicting age-8 outcomes. Thus, we did not address the question of concurrentlinks among variables at age 8. Although we would expect relative stability in set-ting conditions and mother-child relationship factors between ages 4 and 8, it isclear that generalizing the present results to hypothesized concurrent links amongfamily- and peer domains at age 8 would be speculative.

To summarize, we began this study with the idea that attachment securitywould be more important than setting conditions or maternal interactive style inpredicting children's social and emotional adjustment. The results of this studypoint to the importance of both aspects of the mother-child relationship, but alsoindicate that the nature of family and peer links may vary depending upon thesf«cific social behaviors and contexts within which these behaviors are assessed.As a field, it is time to move beyond the simple demonstration of family and peerlinks, and to work toward an elucidation of the processes linking the quality offamily and peer relationships.

Note

1. Note that solitary active and solitary passive behaviors were not included in the age-4 data, becausetheir meaning is less clear in the dyadic context at this age than in the quartet at age 8.

References

Achenbach, T.M. (1991). Manual for the Child Behavior Checklist/4-18.Ainsworth, M. Blehar, M., Waters, E. & Wall, S. (1978). Patterns of attachment: A psycho-

logical study ofthe Strange Situation. Hillsdale, N.J.: Erlbaum.Beck, A.T. (1970). Depression: Causes and treatment. Philadelphia: Univ. of Pennsylvania

Press.Booth, C. & Perman, K. (1989). Attachment Rating Scales. Unpublished manual.Booth, C , Rose-Krasnor, L. & Rubin, K. (1991). Relating preschoolers' social competence

and their mothers' parenting behaviors to early attachment security and high risk status.Journal of Social and Personal Relationships, 8, 363-382.

Bowlby, J. (1973). Attachment and loss. Vol. 2: Separation. New York: Basic Books.Brandt, P.A. & Weinert, C. (1981). The PRQ - a social support measure. Nursing

Research, 30, 277-280.Bretherton, I. (1985). Attachment theory: Retrospect and prospect. In I. Bretherton &

E. Waters (Eds.), Growing points of attachment theory and research. Monographs of theSociety for Research in Child Development, 50 (1-2), Serial No. 209.

Cassidy, J., Marvin, R. & MacArthur Working Group. (1989). Attachment organization inthree and four year olds: Coding guidelines. Unpublished paper. University of Virginia,Charlottesville, VA.

Cohn, D. (1990). Child-mother attachment of six-year-olds and social competence atschool. Child Development, 61, 152-162.

Coplan, R., Rubin, K.H., Fox, N., Calkins, S. & Stewart, S. (1994). Being alone, playingalone and acting alone: Distinguishing among reticence and passive- and active-solitude in young children. Child Development, 65, 129-137.

© BasU Blackwell Ltd. 1994 Social Envelopment, 3, 3, 1994


Cmic, K. & Greenberg, M. (1990). Minor parenting stresses with young children ChildDevelopment, 61, 1628-1637.

Crockenberg, S. (1981). Infant irritability, mother responsiveness and social support influ-ences on the security of infant-mother attachment. Child Development, 51, 857-865.

Dodge, K.A., Bates, J.E. & Pettit, G.S. (1990). Mechanisms in the cycle of violenceScience, 25, 1678-1683.

Erickson, M.F., Sroufe, L.A. & Egeland, B. (1985). The relationship between quality ofattachment and behavior problems in preschool in a high risk sample. In I. Bretherton &E. Waters (Eds.), Growing points of attachment theory and research. Monographs of theSociety for Research in Child Development, 50 (1-2), Serial No., 209.

Greenberg, M., Speltz, M. & DeKlyen, M. (1993). The role of attachment in the earlydevelopment of disruptive behavior pattems. Development and Psychopathology, 5,191-214.

HoUingshead, A.B. (1975). Four Factor Index of Social Status. Unpublished manual.Ladd, G.W. (1991). Family-peer relations during childhood: Pathways to competence and

pathology? Journal of Social and Personal Relationships, 8, 307-314.Lewis, M., Feiring, C , McGuffog, C. & Jaskir, J. (1984). Predicting psychopathology in

six-year-olds from early social relations. Child Development, 55, 123-136.Lieberman, A. (1977). Preschoolers' competence with a peer: Relations with attachment

and peer experience. Child Development, 48, 1277-1287.Maccoby, E. & Martin, J. (1983). Socialization in the context of the family. In E.

Hetherington (Ed.), Handbook of" child psychology, Vol 4: Socialization, personality andsocial development, pp. 1-101. New York: Wiley.

Main, M., Kaplan, N. & Cassidy, J. (1985). Security in infancy, childhood and adulthood:A move to the level of representation. In I. Bretherton & E. Waters (Eds.), Growingpoints of attachment theory and research. Monographs of the Society for Research inChild Development, 50 (1-2), Serial No. 209.

McKinnon, J., Rubin, K., Booth, C. & Rose-Krasnor, L. (1993, March). The influence ofsocioecological and personal support factors on maternal warmth: Mediating effects onattachment security in the preschool years. Poster presented at the Biennial Meeting ofthe Society for Research in Child Development, New Orleans.

Mills, R. & Rubin, K.H. (1993). Socialization factors in the development of social with-drawal. In K. Rubin & J. Asendorpf (Eds.), Social withdrawal, inhibition, and shyness,pp. 117-150. Hillsdale, NJ: LEA.

Park, K.A. & Waters, E. (1989). Security of attachment and preschool friendships. ChildDevelopment, 60, 1076-1081.

Parke, R.D., MacDonald, K.B., Beitel, A. & Bhavnagri, N. (1988). The role of the familyin the development of peer relationships. In R.D. Peters & R.J. McMahan (eds.),Marriages and families: Behavioral treatments and processes, pp. 17^M. New York:Brunner-Mazel.

Parker, J. & Asher, S. (1987). Peer acceptance and later personal adjustment: Are low-accepted children "at risk"? Psychological Bulletin, 102, 357-389.

Pastor, D.L. (1981). The quality of mother-infant attachment and its relationship to tod-dler's initial sociability with peers. Developmental Psychology, 17, 326-335.

Pepler, D. & Rubin, K.H. (Eds.). (1991). The development and treatment of childhoodaggression. Hillsdale, NJ: LEA.

Putallaz, M. (1987). Matemal behavior and children's sociometric status. ChildDevelopment, 58, 324-340.

Radke-Yarrow, M., Cummings, E., Kuczynski, L. & Chapman, M. (1985). Patterns ofattachment in two- and three-year-olds in normal families and families with parentaldepression. Child Development, 56, 884-893.

Rose-Krasnor, L., Rubin, K. & Booth, C. (1993, March). Maternal control, mother-childattachment and preschoolers' nonsocial play. Paper presented at the Biennial Meeting ofthe Society for Research in Child Development, New Orleans, (1993).

Rose-Krasnor, L., Rubin, K.H., Booth, C.L. & Coplan, R. (1994). Maternal directivenessand child attachment security as predictors of social competence in preschoolers.Manuscript submitted for publication.

Rubin, K.H. (1989). The Play Observation Scale (POS). Unpublished manual.

© Basil BiackweU Ltd. 1994 Social Development, 3, 3, 1994


Rubin, K.H. & Asendorpf, J. (1993). Social withdrawal, inhibition and shyness in child-hood: Conceptual and definitional issues. In K.H. Rubin, & J. Asendorpf (Eds.) Socialwithdrawal, inhibition, and shyness in childhood, pp. 3-18. Hillsdale, NJ: LEA.

Rubin, K.H. & McKinnon, J.E. (1993). Maternal Warmth and Control Rating Scales.Unpublished manual.

Sarason, I., Johnson, H. & Siegel, M (1978). Assessing the impact of life changes:Development of the life experiences survey. Journal of Consulting and ClinicalPsychology, 46, 932-946.

Spieker, S. & Booth, C. (1988). Maternal antecedents of attachment quality. In J. Belsky &T. Nezworski (Eds.), Clinical applications of attachment, pp. 95-136. Hillsdale, NJ: LEA.

Sroufe, A. (1983). Infant-caregiver attachment and patterns of adaptation in preschool:The roots of maladaption and competence. In M. Perlmutter (Ed.), MinnesotaSymposium on Child Psychology, Vol. 16, pp. 41-91. Hillsdale, NJ: LEA.

Sroufe, A. & Waters, E (1977). Attachment as an organizational construct. ChildDevelopment, 48, 1184^1119.

Teti, D.M., Nakagawa, M., Das, R. & Wirth, O. (1991). Security of attachment betweenpreschoolers and their mothers: Relations among social interaction, parenting stress, andmothers' sorts of the Attachment Q-set. Developmental Psychology, 27, 440-447.

Turner, P.J. (1991). Relations between attachment, gender, and behavior with peers inpreschool Child Development, 62, 1475-1488.

Waters, E., Posada, G., Crowell, J. & Lay, K. (1993). Is attachment theory ready to con-tribute to our understanding of disruptive behavior problems? Development andPsychopathology, 5, 215-224.

Weinraub, M. & Wolf, B. (1987). Stressful life events, social supports and parent-childinteraction: Similarities and differences in single and two parent families. In Z. Boukydis(Ed.), Research on support for parents and infants in the post natal period, pp. 114-133.Norwood, NJ: Ablex.

Weiss, R. (1974). The provision of social relationships. In Z. Rubin (Ed.), Doing unto oth-ers, pp. 17-26. Englewood Cliffs, NJ: Prentice-Hall.

Youngblade, L.M. & Belsky, J. (1992). Parent-child antecedents of 5-year-olds' closefriendships: A longitudinal analysis. Developmental Psychology, 28, 700-713.

© Basa Blaekwell Ltd. 1994 Social Development, 3, 3, 1994

Predicting social adjustment in middle childhood: the role...

Documents

Transcript of Predicting social adjustment in middle childhood: the role...